Graph-to-text generation on WebNLG seen (test)
65.3BLEU ScoreGPT-2-large
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-2-largeLearning Paradigm=Supervised, Training Setup=Synthetic corpus (934k samples), Approach Type=End-to-end2022.12 | 65.3 | 46 | |
| T5-largeLearning Paradigm=Supervised, Training Setup=Synthetic corpus (934k samples), Approach Type=End-to-end2022.12 | 64.9 | 46 | |
| MELBOURNELearning Paradigm=Supervised, Training Setup=7k in-domain samples, Approach Type=End-to-end2022.12 | 54.5 | 41 | |
| MURMURLearning Paradigm=Few-shot, Number of demonstrations (k)=12022.12 | 41.4 | 37.1 | |
| Direct PromptingLearning Paradigm=Few-shot, Number of demonstrations (k)=52022.12 | 39.9 | 34.3 | |
| Direct PromptingLearning Paradigm=Few-shot, Number of demonstrations (k)=12022.12 | 33.1 | 30.4 | |
| CoT PromptingLearning Paradigm=Few-shot, Number of demonstrations (k)=12022.12 | 22.2 | 22.3 |